Kling 3.0 Image-to-Video API

Generate video from a first frame or first-and-last frames plus prompt and optional Elements with Kling 3.0. Async task API: submit, poll, then download.

First frame: one first-frame image + optional text / Elements
First & last frames: first-frame image + last-frame image + optional text / Elements

Supports multi-shot storytelling, native audio, 720p / 1080p / 4k resolution, and 3–15 second duration.

This is an async task-based API: submit a request to receive a task id, poll for status, then download the video when complete.

Only Kling 3.0 is supported.

Note: First-frame and first-and-last-frame modes are supported. Last-frame-only generation is not supported.

Base URL

https://api.autorouter.top — Production

Authentication

BearerAuth: http (bearer) Authenticate using a Bearer Token. Format: Authorization: Bearer sk-xxxxxx

Endpoints

POST `/kling/image-to-video/kling-3.0`

Create Video Generation Task

Create a video generation task from a first frame or first-and-last frames.

Request Body

Content-Type: application/json

Example (first frame):

{
  "contents": [
    {
      "type": "prompt",
      "text": "A girl sat on the train, looking out the window with a melancholic expression, her head swaying with the train."
    },
    {
      "type": "first_frame",
      "url": "https://example.com/first.png"
    }
  ],
  "settings": {
    "resolution": "4k",
    "duration": 10,
    "audio": "off",
    "multi_shot": false
  },
  "options": {
    "callback_url": "https://xxx/callback",
    "external_task_id": "",
    "watermark_info": {
      "enabled": false
    }
  }
}

Example (first & last frames + element):

{
  "contents": [
    {
      "type": "prompt",
      "text": "A girl sat on the train, looking out the window with a melancholic expression, her head swaying with the train."
    },
    {
      "type": "first_frame",
      "url": "https://example.com/first.png"
    },
    {
      "type": "last_frame",
      "url": "https://example.com/last.png"
    },
    {
      "type": "element",
      "element_id": "163",
      "id": "element_1"
    }
  ],
  "settings": {
    "resolution": "1080p",
    "duration": 10,
    "audio": "native",
    "multi_shot": false
  },
  "options": {
    "callback_url": "https://xxx/callback",
    "external_task_id": "",
    "watermark_info": {
      "enabled": true
    }
  }
}

Properties:

Name	Type	Required	Default	Enum	Description
`contents`	object[]	Yes	-	-	Reference inputs (prompts, images, Elements, etc.)
`contents[].type`	string	Yes	-	`prompt`, `first_frame`, `last_frame`, `element`	Input type
`settings`	object	No	-	-	Output config such as resolution and duration
`settings.multi_shot`	boolean	No	`true`	-	Whether to generate multi-shot video
`settings.audio`	string	No	`off`	`native`, `off`	Whether to generate audio
`settings.resolution`	string	No	`720p`	`720p`, `1080p`, `4k`	Video resolution
`settings.duration`	integer	No	`5`	`3`–`15`	Duration in seconds
`options`	object	No	-	-	General config such as callback URL and watermark
`options.callback_url`	string	No	-	-	Callback URL for task status changes
`options.external_task_id`	string	No	-	-	Custom task ID; must be unique within the account
`options.watermark_info`	object	No	-	-	Whether to also generate a watermarked result

contentsobject[](Required)

Reference input collection. Example format:

[
  {
    "type": "prompt",
    "text": "string"
  },
  {
    "type": "first_frame",
    "url": "string"
  },
  {
    "type": "last_frame",
    "url": "string"
  },
  {
    "type": "element",
    "element_id": "string",
    "id": "string"
  }
]

typestring(Required)

Input type:

prompt: Text prompt
first_frame: First-frame image
last_frame: Last-frame image
element: Element

type=promptobject(Mode-dependent)

{
  "type": "prompt",
  "text": "string"
}

text: Prompt content, max 3072 characters (recommended ≤ 2500)
Multi-shot format: "shot n, m, words; shot n, m, words;" (separated by standard semicolons), where:
- n: shot sequence number (1–6 shots supported)
- m: shot duration in seconds (each shot ≥ 1s; sum must equal total video duration)
- words: shot prompt (max 512 characters)
Reference an Element in the prompt with @xxx, e.g. @Zhang
Avoid Element names that are substrings of each other (e.g. @Zhang and @ZhangSan)
Avoid Element names that overlap with prompt content (e.g. do not use @gmail if the prompt contains wang@gmail.com)

type=first_frame / last_frameobject(Mode-dependent)

{
  "type": "first_frame",
  "url": "string"
}

first_frame is required; last_frame is optional
url: Image URL or Base64
Formats: .jpg, .jpeg, .png
Size: max 50MB
Dimensions: width and height ≥ 300px; aspect ratio between 1:2.5 and 2.5:1
Supports first-frame and first-and-last-frame generation; last-frame-only is not supported

type=elementobject(Optional)

{
  "type": "element",
  "element_id": "string",
  "id": "string"
}

element_id: Element ID returned by the Element-related API
id: Input index ID used in the prompt (e.g. @element_1); must be unique within the same task
At most 3 Elements per task

settingsobject(Optional)

Output configuration such as resolution and duration.

multi_shotboolean(Optional)

Whether to generate multi-shot video. Default true.

When set to false, multi-shot prompts will not produce multi-shot output.

audiostring(Optional)

Whether to generate audio for the video. Default off.

native: Include native audio matching the visuals
off: No audio

resolutionstring(Optional)

Output video resolution. Default 720p.

720p: 720P output
1080p: 1080P output
4k: 4K output

durationinteger(Optional)

Video duration in seconds. Default 5.

Options: 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15

optionsobject(Optional)

General configurations such as callback URL and watermark.

{
  "callback_url": "https://example.com/cb",
  "external_task_id": "string",
  "watermark_info": {
    "enabled": false
  }
}

callback_urlstring(Optional)

Callback URL for task result notifications. If set, the server sends a notification when the task status changes.

external_task_idstring(Optional)

Custom task ID. Does not overwrite the system-generated task ID and can be used for querying. Must be unique within the account.

watermark_infoobject(Optional)

Whether to also generate a watermarked result, controlled by enabled:

{
  "enabled": false
}

true: Include watermarked result
false (default): No watermark

Custom watermarks are not supported.

Responses

200: Task created successfully

Content-Type: application/json

Request Example

curl -X POST "https://api.autorouter.top/kling/image-to-video/kling-3.0" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxx" \
  -d '{
  "contents": [
    {
      "type": "prompt",
      "text": "A girl sat on the train, looking out the window with a melancholic expression, her head swaying with the train."
    },
    {
      "type": "first_frame",
      "url": "https://example.com/first.png"
    }
  ],
  "settings": {
    "resolution": "4k",
    "duration": 10,
    "audio": "off",
    "multi_shot": false
  }
}'

Response Example

{
  "code": 0,
  "message": "string",
  "request_id": "string",
  "data": {
    "id": "string",
    "status": "submitted",
    "create_time": 1781080778802,
    "update_time": 1781080794151,
    "external_id": "string"
  }
}

Response fields:

Name	Type	Description
`code`	integer	Error code; `0` means success
`message`	string	Error message
`request_id`	string	Request ID generated by the system
`data.id`	string	System-generated task ID
`data.status`	string	Task status: `submitted`, `processing`, `succeeded`, `failed`
`data.create_time`	integer	Task creation time, Unix timestamp (ms)
`data.update_time`	integer	Task update time, Unix timestamp (ms)
`data.external_id`	string	Custom task ID, if any

GET `/kling/tasks`

Query Video Generation Task

Query async task status and results by task ID or custom task ID.

Use either task_ids or external_task_ids, not both
Batch queries are supported; separate multiple IDs with ,
Generated results are cleared after about 30 days; save them promptly

Query Params

Name	Type	Required	Description
`task_ids`	string	No	System-generated task IDs, comma-separated
`external_task_ids`	string	No	Custom task IDs, comma-separated

Responses

200: Task queried successfully

Content-Type: application/json

Request Example

curl -X GET "https://api.autorouter.top/kling/tasks?task_ids={id}" \
  -H "Authorization: Bearer sk-xxxxxx"

Response Example

{
  "code": 0,
  "message": "string",
  "request_id": "string",
  "data": [
    {
      "id": "893605946402811985",
      "status": "succeeded",
      "message": "string",
      "create_time": 1781080778802,
      "update_time": 1781080794151,
      "external_id": "string",
      "outputs": [
        {
          "type": "video",
          "id": "string",
          "url": "string",
          "watermark_url": "string",
          "duration": "10"
        }
      ],
      "billing": [
        {
          "charge_type": "cash",
          "cash_type": "balance",
          "amount": "string",
          "package_type": "video",
          "list_price": "string"
        }
      ]
    }
  ]
}

Response fields:

Name	Type	Description
`code`	integer	Error code; `0` means success
`message`	string	Error message
`request_id`	string	Request ID generated by the system, for tracing and troubleshooting
`data`	object[]	Task list
`data[].id`	string	Task ID being queried
`data[].status`	string	Task status: `submitted`, `processing`, `succeeded`, `failed`
`data[].message`	string	Status message; failure reason when failed (e.g. content moderation)
`data[].create_time`	integer	Task creation time, Unix timestamp (ms)
`data[].update_time`	integer	Task update time, Unix timestamp (ms)
`data[].external_id`	string	Custom task ID, if any
`data[].outputs`	object[]	Generated results; fields vary by `type`
`data[].billing`	object[]	Billing details

statusstring

Task status:

submitted: Submitted
processing: Processing
succeeded: Succeeded
failed: Failed

Recommended poll interval: 3–5 seconds until status is succeeded or failed.

outputsobject[]

Generated results. Fields depend on type. Possible values: image, video, audio, element, voice. Image-to-video success typically returns type=video.

type=videoobject

Name	Type	Description
`type`	string	Fixed as `video`
`id`	string	Video ID generated by the system
`url`	string	Result URL (hotlink-protected; cleared after ~30 days; save promptly)
`watermark_url`	string	Watermarked result URL (hotlink-protected)
`duration`	string	Generated video duration in seconds

type=imageobject

Name	Type	Description
`type`	string	Fixed as `image`
`url`	string	Result URL (hotlink-protected; cleared after ~30 days)
`watermark_url`	string	Watermarked image download URL (hotlink-protected)
`group_id`	string	Only present for grouped images; marks group relationship

type=audioobject

Name	Type	Description
`type`	string	Fixed as `audio`
`id`	string	Audio ID generated by the system
`mp3_url`	string	MP3 result URL (hotlink-protected; cleared after ~30 days)
`wav_url`	string	WAV result URL (hotlink-protected; cleared after ~30 days)
`mp3_duration`	string	MP3 duration in seconds
`wav_duration`	string	WAV duration in seconds

type=voiceobject

Name	Type	Description
`type`	string	Fixed as `voice`
`id`	string	Audio ID generated by the system
`name`	string	Audio name
`url`	string	Material download link
`owned_by`	string	Voice source: `kling` for official library, numbers for creator ID
`status`	string	Status: `succeeded`, `deleted`

type=elementobject

Name	Type	Description
`type`	string	Fixed as `element`
`id`	string	Element ID generated by the system
`name`	string	Element name
`description`	string	Element description
`element_type`	string	Element type: `video_character_elements`, `multi_image_elements`
`references`	object[]	Related materials
`references[].type`	string	Material type: `image`, `video`, `voice`
`references[].role`	string	Material role; image: `frontal` / `reference`; video and voice: `refer`
`references[].url`	string	Material download link
`references[].id`	string	Voice ID (when `type=voice`)
`references[].name`	string	Voice name (when `type=voice`)
`references[].owned_by`	string	Voice source (when `type=voice`)
`owned_by`	string	Element source: `kling` for official library, numbers for creator ID
`status`	string	Status: `succeeded`, `deleted`
`tags`	object[]	Element tags
`tags[].id`	integer	Tag ID
`tags[].name`	string	Tag name
`tags[].description`	string	Tag description

billingobject[]

Billing details.

charge_typestring

Account charge type:

cash: Balance deduction
unit: Resource package deduction

cash_typestring

Balance type; only present when charge_type=cash:

balance: Official quota
test_balance: Test quota

amountstring

Deduction amount. When charge_type=cash, the discounted balance price; when charge_type=unit, the resource package units deducted (decimal string).

package_typestring

Resource package type; only present when charge_type=unit. Enum: video, image, audio.

list_pricestring

List price for balance deduction; only present when charge_type=cash.

Error Handling

HTTP 400 Parameter Errors

Scenario	Suggestion
Missing first-frame image	Provide at least one `type=first_frame`
Last frame only	Not supported; include a first frame
More than 3 Elements	Use at most 3 Elements
Invalid image format / size	Follow `first_frame` / `last_frame` limits
Invalid `duration` / `resolution`	Use documented enum values
Multi-shot duration mismatch	Ensure sum of shot durations equals `settings.duration`

HTTP 401 / 403 Auth Errors

401 Unauthorized: Invalid or expired API Key
403 Forbidden: API Key is not allowed to access this model

HTTP 402 Insufficient Balance

Insufficient balance. Please top up in the AutoRouter console.

Task `failed` Status

Cause	Suggestion
Content moderation failed	Adjust the prompt or change the reference image
Image URL unreachable	Ensure public access, or use Base64
Invalid parameter combination	Follow the Request Body rules

When a task fails, AutoRouter automatically refunds your account.

Image-to-Video

contentsobject[](Required)

Properties

settingsobject(Optional)

Properties

optionsobject(Optional)

Properties

statusstring

outputsobject[]

Properties

billingobject[]

Properties

On this page